Review
TRENDS in Biochemical Sciences
541
Vol.28 No.10 October 2003
More than folding: localized functions
of cytosolic chaperones
Jason C. Young, José M. Barral and F. Ulrich Hartl
Cellular Biochemistry, Max Planck Institute of Biochemistry, Am Klopferspitz 18a, D-82152 Martinsried, Germany
Compared with other chaperone systems, heat shock
proteins Hsp70 and Hsp90 interact with a larger variety
of co-chaperone proteins that regulate their activity or
aid in the folding of specific substrate proteins.
Although many co-chaperones are soluble cytosolic
proteins, co-chaperone domains are also found in modular adaptor proteins, which are often localized to intracellular membranes or elements of the cytoskeleton.
These specialized co-chaperones include auxilin,
cysteine string protein, Tom70, UNC-45 and homologs
of Bag-1. The localized co-chaperones can harness the
ATP-dependent mechanisms of Hsp70 and Hsp90 to do
conformational work in diverse functional contexts,
including vesicle secretion and recycling, protein transport and the regulated assembly and/or disassembly of
protein complexes. Such flexibility is unique to the
cytosolic Hsp70 and Hsp90 chaperone system.
It is generally accepted that molecular chaperones interact
with folding intermediates of polypeptides, preventing
non-productive interactions that would result in aggregation, and aiding them in reaching their native state. The
chaperones belonging to the heat shock protein Hsp70,
Hsp90 and chaperonin families act through cycles of
substrate binding and release governed by ATP binding
and hydrolysis [1 –3]. In the eukaryotic cytosol, Hsp70 and
Hsc70 (70-kDa heat shock protein and cognate protein)
handle a broad range of substrate polypeptides, and
contribute generally to the folding of newly synthesized
proteins and refolding of proteins after stress denaturation
[2]. Hsp90 (90-kDa heat shock protein) and the chaperonin
CCT/TRiC are thought to mediate folding of a more limited
set of substrates, including various signal-transducing
proteins in the case of Hsp90 [2,3]. Because Hsp90 often
functions together with Hsc70, the two chaperones might
be considered as parts of a larger multi-chaperone system.
Cytosolic Hsc70 and Hsp90 are further distinguished by
the large number of regulatory or accessory co-chaperone
proteins that they interact with. Many of these cochaperones have a modular architecture in which a
chaperone-interacting domain is fused to other sequences
supplying different activities. Interestingly, several of
these co-chaperone modules are targeted within the
cytosolic compartment to different membrane systems or
to cytoskeletal elements. Such localized co-chaperones can
then recruit cytosolic Hsc70 or Hsp90 for tasks involving
Corresponding author: Jason C. Young ([email protected]).
the conformational modulation of specific target proteins
at the respective intracellular sites. Recent work has
revealed some of the mechanisms and cellular processes
supported by the targeted co-chaperones and their
cytosolic chaperone partners. This review will focus on
the best characterized of these specific co-chaperones,
acting in endocytosis and exocytosis, protein targeting,
cytoskeletal function and signal transduction.
Three classes of co-chaperone domains appear as
protein modules: DnaJ homology or J domains, Bag-1
homology or Bag domains, and so-called tetratricopeptide
repeat (TPR) clamp domains. Representative structures of
these domains are depicted in Figure 1 [4– 6]. DnaJ
homologs were the first co-chaperone proteins to be
recognized as a family, and contain a conserved His-ProAsp tripeptide motif that is essential for function [7]. The
J domains of these homologs stimulate ATP hydrolysis via
the partner Hsc70 proteins, converting them to the ADPbound forms and leading to stable binding of substrates by
the Hsc70 peptide-binding domains [1,2]. More recently,
Bag-1 homologs were also identified as a family of Hsc70
cofactors [8]. Bag domains interact with the ATPase
(a) J domain
(c)
C
Bag domain
N
HPD motif
(b) TPR clamp domain
Hsc70 ATPase
domain
Hsp90 C-terminal peptide
Ti BS
Figure 1. Molecular structures of modular co-chaperone domains. (a) Nuclear magnetic resonance structure of the human Hsp40 J domain (red), the position of the
conserved His-Pro-Asp (HPD) tripeptide motif necessary for function is marked. (b)
Crystallographic structure of the central tetratricopeptide (TPR) clamp domain of
human Hop (light blue) bound to the C-terminal peptide of Hsp90 (black ball and
stick); the N-terminal TPR clamp domain of Hop binds the Hsc70 C-terminal peptide similarly. (c) Crystallographic structure of the Bag domain of human Bag-1
(yellow) bound to the Hsc70 ATPase domain (light green). Ribbon diagrams were
generated with MOLSCRIPT and RASTER-3D.
http://tibs.trends.com 0968-0004/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2003.08.009
542
Review
TRENDS in Biochemical Sciences
Vol.28 No.10 October 2003
Source
Chaperone
partner
Specialized
domains
Localization
Function
Animals
Hsc70
Clathrin binding,
kinase
CCV,
PM
Clathrin uncoating,
vesicle budding
Swa2/Aux1
S. cerevisiae
Hsc70
Clathrin binding
CCV
Clathrin uncoating,
ER inheritance
CSP
Animals
Hsc70
(Hsp90)
Acylation,
SGT binding
SV
Exocytosis
SGT
Animals
Hsc70
CSP binding
cytosol,
SV
Exocytosis
Djp1
S. cerevisiae
Hsc70
?
Cytosol,
(PER?)
Peroxisomal
targeting
Tom70
Animals, fungi
Hsc70,
Hsp90
TM,
pre-protein binding
MOM
Mitochondrial
import
Tom34
Mammals
Hsp90
?
Cytosol,
(MOM)
Mitochondrial
targeting?
UNC-45
Animals
Hsp90
Myosin binding
SARC,
cytosol
Myosin complex
folding, assembly
Mammals
Hsp90
PPIase,
dynein binding
MT,
cytosol
Nuclear
transport?
Mrj
Mammals
Hsc70
K8/18 binding
IF
IF organization
SODD/Bag-4
Mammals
Hsc70
?
PM,
cytosol
TNFR-1
inactivation
CAIR-1/Bag-3
Mammals
Hsc70
PLCγ binding
PM,
cytosol
PLCγ signaling
Snl1
S. cerevisiae
Hsc70
TM
ER, NM
NP biogenesis?
Name
auxilin
GAK
FKBP52
Cyp40
Ti BS
Figure 2. Schematic diagram of localized co-chaperone proteins. N termini are on the left, J domains are shown in red, tetratricopeptide clamp domains in light blue, and
Bag domains in yellow. Specialized interaction domains are depicted as follows: clathrin-binding domains of auxilin, GAK and Swa2/Aux1, dark gray; kinase domain of
GAK, light gray; acylated cysteines of CSP, light orange; interaction sites between CSP and SGT, green; transmembrane domains of Tom70 and Snl1, black; preproteinbinding domain of Tom70, light brown; myosin-binding domain of UNC-45, magenta; PPIase- and dynein-binding domains of FKBP52 and Cyp40, pink; K8/18-binding
domain of Mrj, dark blue; PLCg-binding site of CAIR-1/Bag-3, light green. Abbreviations: CCV, clathrin-coated vesicles; ER, endoplasmic reticulum; GAK, cyclin G-associated
kinase; IF, intermediate filaments; K8/18, keratin 8 and 18 filaments; MOM, mitochondrial outer membrane; MT, microtubules; NM, nuclear membrane; NP, nuclear pore;
PER, peroxisomes; PLCg, phospholipase C-g; PM, plasma membrane; PPIase, peptidylprolyl isomerase; SARC, sarcomere; SODD, suppressor of death domains; SV, synaptic vesicles; TM, transmembrane domain; TNFR-1, tumor necrosis factor receptor type 1; TPR, tetratricopeptide repeat.
domains of eukaryotic cytosolic Hsc70 and trigger the
exchange of ADP for ATP, favoring release of Hsc70-bound
peptides [5,9]. TPR clamp domains were first identified as
homologous sequences in a series of cytosolic Hsp90 cochaperones [10], and are now known to recognize the
C-terminal peptides of Hsp90 or Hsc70, or both [3,6]. At
least one TPR clamp protein, Hop/Sti1, inhibits the
ATPase cycle of Hsp90, whereas other such co-chaperones
do not [11]. Thus, the localized co-chaperones of Hsc70 and
Hsp90 can control the biochemical regulation as well as
the intracellular location of the recruited chaperones.
An important comparison can be made to the action of
organellar Hsp70s in polypeptide transport across membranes. The unique mitochondrial-inner-membrane cochaperone Tim44 localizes mitochondrial matrix Hsp70
(mtHsp70/Ssc1) to drive import into mitochondria [12],
and Bip/Kar2 (ER lumenal Hsp70) similarly requires the
membrane-anchored J domain of Sec63 to act in ER
http://tibs.trends.com
translocation [13]. Analogously, cytosolic Hsc70 and/or
Hsp90 can be recruited by their localized co-chaperones to
accomplish specific tasks.
Clathrin uncoating
Perhaps the best-known specialized co-chaperone is the
J-domain protein auxilin, which functions in the Hsc70mediated uncoating of clathrin-coated vesicles (CCVs)
budded from the plasma membrane [14]. Neuronal auxilin
contains a central clathrin-binding domain and a J domain
at the extreme C terminus (Figure 2). In vitro studies have
established that the clathrin-binding domain first assembles onto clathrin cages, and the J domain then stimulates
free Hsc70 to hydrolyze ATP. Hsc70 in the ADP state binds
tightly to clathrin, presumably in a manner similar to the
binding of an unfolded polypeptide, and this distorts the
conformation of clathrin leading to disassembly of the cage
[14,15] (Figure 3). In the cytosol, Hsc70 continuously
Review
TRENDS in Biochemical Sciences
PM
Cytosol
Clathrin-coated
vesicle
Ti BS
Figure 3. Clathrin uncoating by auxilin and Hsc70. Auxilin targets via its clathrinbinding domain (gray) to clathrin (dark purple) on the surface of vesicles. The
J domain of auxilin (red) triggers ATP hydrolysis by Hsc70 (light purple), which
then binds clathrin tightly. Hsc70 binding causes dissociation of clathrin from the
vesicle surface, and enables it to recycle onto membranes.
releases and re-binds clathrin, and together with clathrin
assembly proteins such as AP-180, stabilizes clathrin for
subsequent re-assembly on the plasma membrane [16].
The non-neuronal form of auxilin, called GAK (cyclin
G-associated kinase) or auxilin 2, contains an
additional N-terminal kinase domain (Figure 2) that
is not required for clathrin uncoating but might
regulate clathrin assembly [17,18].
Different experimental approaches have confirmed the
functions of Hsc70 and auxilin in clathrin uncoating in
vivo: (i) by overexpressing dominant inhibitory mutants of
Hsc70 in cultured HeLa cells; (ii) by RNA interference
knockdown of auxilin expression in Caenorhabditis
elegans; (iii) by injection of mutant auxilin lacking a
functional J domain into squid presynaptic nerve terminals; and (iii) by a genetic screen in Drosophila for
endocytosis mutants revealing a point mutation in Hsc70
[19 – 22]. In all cases, clathrin-dependent endocytosis is
markedly inhibited and is accompanied by abnormal
vesicular accumulation of clathrin. Disruption of Hsc70
function in HeLa cells causes defects not only in
endocytosis, but also in vesicle transport between the
trans-Golgi, endosomes and plasma membrane,
suggesting that Hsc70 together with auxilin or GAK
might be generally required for all clathrin-dependent
vesicle trafficking [19]. Consistent with this, RNA interference of auxilin in C. elegans leads to broad developmental defects and an overall intracellular immobilization
of clathrin [20]. However, a new report suggests that
Hsc70 and auxilin also form complexes with dynamin, a
GTPase that works in the budding and fission of CCVs, and
some of the effects of disrupting auxilin in vivo might arise
from defects in vesicle formation as well as in clathrin
uncoating [23].
The
Saccharomyces
cerevisiae
auxilin
called
Swa2/Aux1 differs from the animal forms in several
interesting aspects. Between the clathrin-binding and
J domains, Swa2 also contains a predicted TPR clamp
domain, which is expected to recognize Hsc70 and/or
Hsp90 [24,25] (Figure 2). The swa2– 1 mutation bears a
single amino-acid exchange in a conserved residue of the
TPR domain. Interestingly, the swa2– 1 mutation shows
synthetic lethality with a deletion of the gene encoding the
ADP-ribosylation factor Arf1 (hence the name Swa2,
http://tibs.trends.com
Vol.28 No.10 October 2003
543
synthetic lethal with arf1), which functions in coatedvesicle transport between the endoplasmic reticulum (ER)
and Golgi. Although deletion of the gene encoding Swa2
causes endocytosis defects, the swa2– 1 point mutation
does not interfere with clathrin binding or stimulation of
the Hsc70 ATPase by the Swa2 J domain [24]. Thus,
chaperone binding to the TPR domain might serve a
function that is unique to yeast auxilin. One possibility is
the involvement of Swa2 in the formation of the ER during
cell division in yeast, which is a function not known for
auxilin or GAK in animal cells [26].
Synaptic vesicle fusion
Cysteine string protein (CSP) is a specialized J-domain cochaperone and was first shown to function in calciumactivated exocytosis of synaptic vesicles [27,28]. CSP has
been identified in Drosophila, Torpedo and mammals (with
some species having different isoforms), but is not found in
S. cerevisiae. CSP contains a J domain at its N terminus,
and a central cysteine-rich sequence anchored to vesicle
membranes by multiple acylations [29] (Figure 2). The
involvement of Hsc70 in CSP-mediated exocytosis in vivo
is now supported by a mutagenesis study in Drosophila,
which produced several alleles of Hsc70 that caused
impaired synaptic transmission at the neuromuscular
junction. Importantly, these particular Hsc70 alleles did
not cause endocytosis or vesicle recycling defects,
suggesting that the phenotype is unrelated to the
clathrin-uncoating activity of Hsc70 [30].
The exact mechanism of CSP activity is still unresolved.
CSP interacts with some neuronal SNARE proteins [31],
which are responsible for vesicle targeting and fusion, and
Hsc70 that is recruited by CSP has been proposed to
stabilize unstructured monomeric SNAREs before formation of the fusion-active hetero-oligomeric SNARE
complex [29]. However, there is also some evidence that
CSP might regulate the activity of the calcium channels
that trigger neurotransmitter release. These channels are
controlled by the heterotrimeric GTP-binding proteins
(G proteins) such that, after dissociation from the GTPbound Ga subunit, the free Gbg subunits bind the channels
and down-regulate their activity. CSP forms complexes
with both Gbg and Ga, and calcium-channel modulation by
CSP overexpression in cultured neuronal cells shows
characteristics of Gbg inhibition [32] (Figure 4).
In some cases, CSP acts with another co-chaperone
called SGT (small glutamine-rich TPR protein). SGT has
an Hsc70-binding TPR clamp domain in its central region
and a glutamine-rich sequence of unknown function near
the C terminus (Figure 2). The N-terminal region of SGT
binds the C terminus of CSP (Figure 2), localizing a
fraction of the cytosolic SGT and Hsc70 to vesicle
membranes in a trimeric complex. Together, CSP and
SGT regulate the ATPase activity of Hsc70, and SGT
overexpression in neurons inhibits neuronal exocytosis
[33]. Although it is not yet clear at which step of membrane
fusion the SGT-containing complex works, the TPR
domain of SGT might transiently anchor Hsc70 to the
membrane to handle the relevant polypeptides.
Among these polypeptides could be the small GTPases
of the Rab protein family, particularly the neuronal
544
Review
TRENDS in Biochemical Sciences
Vol.28 No.10 October 2003
Ca2+
PM
Cytosol
Synaptic
vesicle
Ti BS
Figure 4. Functions of cysteine string protein (CSP) in synaptic vesicle exocytosis. CSP is attached to membranes by its acylated cysteine residues (dark green). CSP interacts through its J domain (red) with Hsc70 (light purple), and with interacts with the Ga (large orange oval) and Gbg subunits (small orange ovals) of heterotrimeric G proteins to modulate calcium channel activation of exocytosis. CSP might also recruit Hsc70 to stabilize monomeric SNARE proteins (dark purple) before SNARE-complex
formation, vesicle fusion and neurotransmitter release (lime green). After vesicle fusion, CSP works with Hsp90 (brown) and Hsc70 to release the regulatory complex of
Rab3a (pink) and aGDI (pale green) from the membrane and recycle Rab3a onto vesicles. Some functions of CSP involve the co-chaperone SGT, which binds to Hsc70
through its TPR clamp domain (blue) and to CSP through its N-terminal domain (light gray rectangle). Abbreviations: aGDI, GDP-dissociation inhibitor; CSP, cysteine string
protein; PM, plasma membrane; TPR, tetratricopeptide repeat.
isoform Rab3A, which limits the extent of calciumstimulated exocytosis. GTP hydrolysis by Rab3A on vesicle
membranes is triggered upon vesicle fusion, and the brainspecific GDP-dissociation inhibitor a (aGDI) then transfers Rab3A from the membrane to the cytosol. It seems
that the aGDI binds Rab3A on the membrane in a complex
containing CSP, Hsc70 and Hsp90. Interestingly, both the
removal of Rab3A from membranes and the subsequent
dissociation of aGDI are blocked by the anti-tumor drugs
geldanamycin and radicicol, which specifically inhibit the
Hsp90 ATPase cycle. The binding of aGDI by Hsp90, and
probably also by Hsc70, is therefore proposed to facilitate
both the membrane extraction of Rab3A and its reassembly on the membrane in the GTP-bound state [34]
(Figure 4). It will be interesting to see if a similar
mechanism also applies to other Rab proteins, or even to
other members of the GTPase superfamily that includes
the heterotrimeric G proteins and Arf1.
Protein targeting
Cytosolic Hsc70 is known to be important for the posttranslational translocation of proteins across the ER,
mitochondrial and peroxisomal membranes [35 – 37]. This
Hsc70 function has been thought to be non-specific for the
target organelles, most probably involving the prevention
of precursor protein aggregation. However, genetic analysis has shown that the S. cerevisiae co-chaperone Djp1 is
specifically involved in targeting to peroxisomes, but not to
the ER or mitochondria. Djp1 has a J domain at its
N terminus, and C-terminal sequences that might interact
with the peroxisomal surface (Figure 2), although only a
fraction of the protein co-localizes with membranes [38].
The molecular mechanism of Djp1 function with Hsc70
http://tibs.trends.com
could thus involve cycling with precursor proteins from the
cytosol to the peroxisomal membrane.
Recently, the mitochondrial import receptor Tom70
(70-kDa translocase of the outer membrane) has also been
shown to act as a localized co-chaperone [39]. From yeast to
mammals, Tom70 functions in the post-translational
import of mitochondrial proteins having non-classical,
internal targeting sequences [40,41]. An N-terminal
domain anchors Tom70 to the cytoplasmic face of the
mitochondrial outer membrane (MOM), followed by a
chaperone-binding TPR clamp domain and another
domain that binds directly to preproteins (Figure 2).
Mammalian Tom70 recognizes both Hsp90 and Hsc70, but
the yeast Tom70 is specific for Hsc70 (the Ssa proteins).
When chaperone binding to the Tom70 TPR domain is
disrupted by competition with an Hsp90 fragment in vitro,
or mutation of Tom70 in yeast, mitochondrial targeting is
inhibited and the oligomeric preprotein-Tom70 complex
that precedes transport across the MOM cannot assemble.
Furthermore, mitochondrial import of Tom70-dependent
preproteins in mammals is blocked by the Hsp90-specific
inhibitor geldanamycin both in vitro and in Cos7 cultured
cells [39], suggesting that the chaperones might also help
transfer preproteins from the Tom70 complex to the
intermembrane space (Figure 5).
The mammalian Tom34 protein, which has no counterpart in yeast, might also function in the targeting of
mitochondrial proteins with classical N-terminal presequences [42]. Tom34 contains a TPR clamp domain
that recognizes Hsp90 [43] (Figure 2). However, Tom34 is
almost entirely cytosolic with only a small fraction
associated with the MOM [42], and possibly acts with
Hsp90 in handling preproteins before they reach the
membrane.
Review
TRENDS in Biochemical Sciences
545
Vol.28 No.10 October 2003
Cytosol
MOM
MIM
Ti BS
Figure 5. Protein targeting via chaperones and Tom70. Hsc70 (light purple) and also Hsp90 (brown) in mammals, bind some mitochondrial preproteins (black) in the cytosol. The chaperones dock onto the TPR clamp domain (blue) of Tom70 on the mitochondrial outer membrane. The preprotein-binding domain of Tom70 (light gray) then
recognizes internal targeting sequences in the preprotein. After ATP cycling by the chaperones, the preprotein is transported by the import machinery (orange) through the
outer membrane. Abbreviations: MIM, mitochondrial inner membrane; MOM, mitochondrial outer membrane; TPR, tetratricopeptide repeat.
Cytoskeleton function
Hsc70 and Hsp90 are recruited not only to intracellular
membranes, but also to the cytoskeletal networks, including motor proteins that move along actin filaments and
microtubules. Again, the recruited chaperones are thought
to fulfill different functions at these macromolecular
assemblies.
In C. elegans, the unc-45 gene product (named for the
uncoordinated phenotype) is required for the correct
assembly of conventional myosin into muscle thick
filaments, which then contact the actin-based thin
filaments of the sarcomere [44,45]. UNC-45 contains an
N-terminal TPR clamp domain and a C-terminal UCS
(UNC-45/Cro1/She4) domain. The TPR clamp binds
specifically to Hsp90, whereas the C-terminal regions
bind and exert chaperone activity on the myosin head
(Figure 2). Hsp90 recruited by UNC-45 probably aids in
the folding and assembly of myosin complexes, and might
also act in their disassembly [46]. Mice and humans each
have two UNC-45 isoforms with the same domain
configuration. The striated muscle (SM) type is expressed
only in cardiac and skeletal muscles and its function might
be restricted to the sarcomere, whereas a general cell (GC)
type of UNC-45 is expressed in multiple tissues and might
have a role in myosin-based cytoskeletal functions [47].
Although fungal UCS domain proteins are also important
for the assembly and function of non-muscle myosins [48],
they do not contain TPR clamp domains. Instead, they
contain N-terminal sequences that might interact with
additional proteins, which in turn might recruit molecular
chaperones.
Intriguingly, TPR-containing co-chaperones also interact with cytoplasmic dynein, a motor protein that moves
along microtubules. These include the immunophilin
FKBP52 (52-kDa FK506-binding protein) and cyclophilin
Cyp40, both of which bind Hsp90 through their TPR clamp
domains. The interaction with dynein is mediated by their
central peptidylprolyl isomerase (PPIase) domains
(Figure 2), but is independent of this enzymatic activity.
These co-chaperones are thought to link Hsp90-bound
http://tibs.trends.com
steroid hormone receptors to the microtubule cytoskeleton
for nuclear import [49,50]. Recent work with S. cerevisiae,
using heterologously expressed glucocorticoid receptor,
suggests that this interaction might not be necessary for
nuclear transport [51], but dynein binding might still
increase the transport efficiency in mammalian cells.
The J domain co-chaperone Mrj (mammalian relative of
DnaJ) has recently been found to connect Hsc70 with the
intermediate filament cytoskeleton. Mrj contains an
N-terminal Hsc70-regulatory J domain, and its C-terminal
domain specifically recognizes the keratin 18 component of
keratin 8 and 18 (K8/18) intermediate filaments (Figure 2).
Notably, microinjection of antibodies against Mrj into
HeLa cells disrupts the K8/18 filament network without
affecting the actin or microtubule cytoskeletons. Thus, it
has been proposed that Hsc70 that has been recruited by
Mrj acts in the regulated assembly of the keratin filament
cytoskeleton [52].
Bag domain co-chaperones
The Bag-1-related family of proteins form another family
of modular co-chaperones [4]. As nucleotide exchange
factors for Hsc70, they probably act by stimulating ATP
cycling by Hsc70 rather than stabilizing its binding to
polypeptides. At least two Bag domain proteins are
associated with plasma membrane proteins (Figure 2):
SODD (silencer of death domains) or Bag-4 with the
cytosolic region of TNFR1 (tumor necrosis factor receptor
type 1), and CAIR-1 (calcium-influx inhibitor response
regulated) or Bag-3 with PLCg (phospholipase C-g). In
both cases, the Bag co-chaperones bind the inactive forms
of these signal transducers, and are released after
activation upon binding of cytokine ligand to TNFR1, or
phosphorylation of PLCg by EGF-R (epidermal growth
factor receptor) [53,54]. The function of Hsc70 in these
pathways is not yet clear, but could be related to
conformational changes required for the switching mechanisms. In S. cerevisiae, the Bag domain co-chaperone
Snl1 (suppressor of Nup116C lethality) is localized to the
cytosolic face of nuclear and ER membranes, and can
546
Review
TRENDS in Biochemical Sciences
suppress growth defects caused by some mutant nuclear
pore proteins, perhaps by stimulating Hsc70 to stabilize
them [55]. The physiological function of Snl1 under normal
growth conditions, however, remains to be established.
Concluding remarks
The list of localized co-chaperones of Hsc70 and Hsp90
continues to grow, and knowledge of how they act will
broaden our understanding of the functions of these
chaperones. Important questions concern how the
recruited Hsc70 or Hsp90 proteins work mechanistically,
and how the chaperone-mediated steps are integrated
with the functions of other components.
A few principles can be derived from what is known
about the chaperone-dependent mechanisms discussed
above. First, the action of the chaperones must be
regulated as well as localized within the cytosol. For
example, there is no evidence that auxilin and Hsc70
remove clathrin from clathrin-coated pits in the plasma
membrane, but only from vesicles after budding [14,15].
Such regulation probably requires interactions in addition
to those between the chaperones and co-chaperones. In the
case of auxilin, it could be speculated that the interaction
with dynamin during the vesicle budding process, but not
afterwards [23], might be such a regulatory interaction.
Another example might be the chaperone-dependent
removal of Rab3a from synaptic vesicle membranes [34]
because the complex with CSP, Hsp90 and Hsp70
recognizes GDP-bound Rab3a together with aGDI, but
probably not GTP-bound Rab3a without aGDI. For other
functions of CSP and Hsp70, including the interaction
with the heterotrimeric G proteins [32], the regulatory
steps remain to be discovered.
Second, additional interactions centered on the cochaperones will also provide further specificity to the
chaperone functions. A clear example is the function of
Tom70: although Hsp90 and Hsc70 help to bring mitochondrial preproteins to the import receptor, sorting of
preproteins from other chaperone substrates must be
performed by Tom70 itself or other components of the
import machinery [39]. The PPIases FKBP52 and Cyp40,
which connect to dynein and to Hsp90, might similarly
contribute to the specificity of microtubule-based
transport because the steroid hormone receptors, but not
other Hsp90-bound substrates, are targeted to the nucleus
[49,50]. It is not known how the dynein-bound complex
distinguishes between liganded receptors, which must be
targeted, and unliganded receptors, which must remain
cytosolic, and the co-chaperones might also act in this
regulatory step. A crucial challenge will be to discover how
the specificity and regulation of chaperone function are
determined by the components of these co-chaperone
complexes.
Third, the chaperones must fulfill their localized
functions through the same biochemical mechanisms
that they use for protein folding. Hsc70 is known to
recognize short hydrophobic polypeptide stretches in an
extended conformation [1,2], and cellular processes that
require Hsc70, such as clathrin uncoating, probably
require the presentation of such chaperone-binding sites.
However, as cytosolic clathrin remains largely folded [16],
http://tibs.trends.com
Vol.28 No.10 October 2003
binding by Hsc70 does not necessarily imply a gross loss of
protein structure. The features recognized by Hsp90 have
not yet been clearly defined, but this chaperone appears to
bind some of its substrates when they are in a compact
state close to their native conformation [3]. This characteristic suggests, for instance, that Hsp90 that has been
recruited to myosin by UNC-45 [45] might act in the final
stages of myosin assembly into thick filaments. The key to
understanding these processes will be to determine which
polypeptides are handled by the chaperones and what
conformational changes are involved.
Fourth, localization of the chaperone partners by their
specialized co-chaperones is most probably necessary to
accomplish the specified task. The localized co-chaperones
must provide a particular orientation, local concentration
or regulatory step of the chaperones that is needed for
their relevant functions. In the absence of a localized cochaperone, binding of chaperones to a crucial site might be
too infrequent, or mechanically ineffective without a
coordinating anchor. The modular domain structure of
the co-chaperones imparts a unique flexibility to the Hsc70
and Hsp90 chaperone system that enables it to act not only
in conventional protein folding, but also in adjusting
protein conformational switches in a multitude of cellular
processes.
Acknowledgements
We thank Andreas Bracher for his help in preparing Figure 1. J.M.B. is
supported by a Human Frontier Science Program fellowship.
References
1 Bukau, B. and Horwich, A. (1998) The Hsp70 and Hsp60 chaperone
machines. Cell 92, 351– 366
2 Hartl, F.U. and Hayer-Hartl, M. (2002) Molecular chaperones in the
cytosol: from nascent chain to folded protein. Science 295, 1852– 1858
3 Young, J.C. et al. (2001) Hsp90: a specialized but essential proteinfolding tool. J. Cell Biol. 154, 267 – 273
4 Qian, Y.Q. et al. (1996) Nuclear magnetic resonance solution structure
of the human Hsp40 (HDJ-1) J-domain. J. Mol. Biol. 260, 224 – 235
5 Sondermann, H. et al. (2001) Structure of a Bag/Hsc70 complex:
convergent functional evolution of Hsp70 nucleotide exchange factors.
Science 291, 1553– 1557
6 Scheufler, C. et al. (2000) Structure of TPR domain-peptide complexes:
critical elements in the assembly of the Hsp70-Hsp90 multichaperone
machine. Cell 101, 199 – 210
7 Caplan, A.J. et al. (1993) Eukaryotic homologues of Escherichia coli
dnaJ: a diverse protein family that functions with HSP70 stress
proteins. Mol. Biol. Cell 4, 555– 563
8 Takayama, S. and Reed, J.C. (2001) Molecular chaperone targeting
and regulation by BAG family proteins. Nat. Cell Biol. 3, E237– E241
9 Höhfeld, J. and Jentsch, S. (1997) GrpE-like regulation of the hsc70
chaperone by the anti-apoptotic protein BAG-1. EMBO J. 16,
6209– 6216
10 Owens-Grillo, J.K. et al. (1996) A model of protein targeting mediated
by immunophilins and other proteins that bind to hsp90 via
tetratricopeptide repeat domains. J. Biol. Chem. 271, 13468 – 13475
11 Prodromou, C. et al. (1999) Regulation of Hsp90 ATPase activity by
tetratricopeptide repeat (TPR)-domain co-chaperones. EMBO J. 18,
754 – 762
12 Bauer, M.F. et al. (2000) Protein translocation into mitochondria: the
role of TIM complexes. Trends Cell Biol. 10, 25 – 31
13 Corsi, A.K. and Schekman, R. (1996) Mechanism of polypeptide
translocation into the endoplasmic reticulum. J. Biol. Chem. 271,
30299 – 30302
14 Ungewickell, E. et al. (1995) Role of auxilin in uncoating clathrincoated vesicles. Nature 378, 632– 635
15 Holstein, S.E.H. et al. (1996) Mechanism of clathrin basket
Review
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
TRENDS in Biochemical Sciences
dissociation: separate functions of protein domains of the DnaJ
homologue Auxilin. J. Cell Biol. 135, 925 – 937
Jiang, R. et al. (2000) Hsc70 chaperones clathrin and primes it to
interact with vesicle membranes. J. Biol. Chem. 275, 8439 – 8447
Greener, T. et al. (2000) Role of cyclin G-associated kinase in uncoating
clathrin-coated vesicles from non-neuronal cells. J. Biol. Chem. 275,
1365 – 1370
Umeda, A. et al. (2000) Identification of the universal cofactor (auxilin
2) in clathrin coat dissociation. Eur. J. Cell Biol. 79, 336 – 342
Newmyer, S.L. and Schmid, S.L. (2001) Dominant-interfering Hsc70
mutants disrupt multiple stages of the clathrin-coated vesicle cycle in
vivo. J. Cell Biol. 152, 607– 620
Greener, T. et al. (2001) Caenorhabditis elegans auxilin: a J-domain
protein essential for clathrin-mediated endocytosis in vivo. Nat. Cell
Biol. 3, 215 – 219
Morgan, J.R. et al. (2001) Uncoating of clathrin-coated vesicles in
presynaptic terminals: roles for Hsc70 and auxilin. Neuron 32,
289 – 300
Chang, H.C. et al. (2002) Hsc70 is required for endocytosis and clathrin
function in Drosophila. J. Cell Biol. 159, 477 – 487
Newmyer, S.L. et al. (2003) Auxilin – dynamin interactions link the
uncoating ATPase chaperone machinery with vesicle formation. Dev.
Cell 4, 929 – 940
Gall, W.E. et al. (2000) The auxilin-like phosphoprotein Swa2p is
required for clathrin function in yeast. Curr. Biol. 10, 1349– 1358
Pishvaee, B. et al. (2000) A yeast DNA J protein required for uncoating
of clathrin-coated vesicles in vivo. Nat. Cell Biol. 2, 958 – 963
Du, Y. et al. (2001) Aux1p/Swa2p is required for cortical endoplasmic
reticulum inheritance in Saccharomyces cerevisiae. Mol. Biol. Cell 12,
2614 – 2628
Zinsmaier, K.E. et al. (1994) Paralysis and early death in cysteine
string protein mutants of Drosophila. Science 263, 977 – 980
Mastrogiacomo, A. et al. (1994) Cysteine string proteins: a potential
link between synaptic vesicles and presynaptic Ca2þ channels. Science
263, 981 – 982
Chamberlain, L.H. and Burgoyne, R.D. (2000) Cysteine-string protein:
the chaperone at the synapse. J. Neurochem. 74, 1781– 1789
Bronk, P. et al. (2001) Drosophila Hsc70-4 is critical for neurotransmitter exocytosis in vivo. Neuron 30, 475– 488
Rothman, J.E. (1994) Mechanisms of intracellular protein transport.
Nature 372, 55 – 63
Magga, J.M. et al. (2000) Cysteine string protein regulates G protein
modulation of N-type calcium channels. Neuron 28, 195 – 204
Tobaben, S. et al. (2001) A trimeric protein complex functions as a
synaptic chaperone machine. Neuron 31, 987 – 999
Sakisaka, T. et al. (2002) Rab-aGDI activity is regulated by a Hsp90
chaperone complex. EMBO J. 21, 6125 – 6135
Deshaies, R.J. et al. (1988) A subfamily of stress proteins facilitates
translocation of secretory and mitochondrial precursor polypeptides.
Nature 332, 800 – 805
Chirico, W.J. et al. (1988) 70K heat shock related proteins stimulate
protein translocation into microsomes. Nature 332, 805 – 810
Walton, P.A. et al. (1994) Involvement of 70-kD heat-shock proteins in
peroxisomal import. J. Cell Biol. 125, 1037– 1046
Vol.28 No.10 October 2003
547
38 Hettema, E.H. et al. (1998) The cytosolic Dnaj-like protein Djp1p is
involved specifically in peroxisomal protein import. J. Cell Biol. 142,
421– 434
39 Young, J.C. et al. (2003) Molecular chaperones Hsp90 and Hsp70
deliver preproteins to the mitochondrial import receptor Tom70. Cell
112, 41 – 50
40 Söllner, T. et al. (1990) A mitochondrial import receptor for the ADP/
ATP carrier. Cell 62, 107 – 115
41 Suzuki, H. et al. (2002) Characterization of rat TOM70 as a receptor of
the preprotein translocase of the mitochondrial outer membrane.
J. Cell Sci. 115, 1895– 1905
42 Chewawiwat, N. et al. (1999) Characterization of the novel mitochondrial protein import component, Tom34, in mammalian cells.
J. Biochem. (Tokyo) 125, 721 – 727
43 Young, J.C. et al. (1998) Specific binding of tetratricopeptide-repeat
proteins to the C-terminal 12 kDa domain of Hsp90. J. Biol. Chem.
273, 18007 – 18010
44 Epstein, H.F. and Thomson, J.N. (1974) Temperature-sensitive
mutation affecting myofilament assembly in Caenorhabditis elegans.
Nature 250, 579 – 580
45 Barral, J.M. et al. (1998) Unc-45 mutations in Caenorhabditis elegans
implicate a CRO1/She4p-like domain in myosin assembly. J. Cell Biol.
143, 1215– 1225
46 Barral, J.M. et al. (2002) Role of the myosin assembly protein UNC-45
as a molecular chaperone for myosin. Science 295, 669 – 671
47 Price, M.G. et al. (2002) Two mammalian UNC-45 isoforms are related
to distinct cytoskeletal and muscle-specific functions. J. Cell Sci. 115,
4013– 4023
48 Wesche, S. et al. (2003) The UCS-domain protein She4p binds to
myosin motor domains and is essential for class I and class V myosin
function. Curr. Biol. 13, 715– 724
49 Galigniana, M.D. et al. (2001) Evidence that the peptidylprolyl
isomerase domain of the hsp90-binding immunophilin FKBP52 is
involved in both dynein interaction and glucocorticoid receptor
movement to the nucleus. J. Biol. Chem. 276, 14884 – 14889
50 Galigniana, M.D. et al. (2002) Binding of hsp90-associated immunophilins to cytoplasmic dynein: direct binding and in vivo evidence that
the peptidylprolyl isomerase domain is a dynein interaction domain.
Biochemistry 41, 13602 – 13610
51 Riggs, D.L. et al. (2003) The Hsp90-binding peptidylprolyl isomerase
FKBP52 potentiates glucocorticoid signaling in vivo. EMBO J. 22,
1158– 1167
52 Izawa, I. et al. (2000) Identification of Mrj, a DnaJ/Hsp40 family
protein, as a keratin 8/18 filament regulatory protein. J. Biol. Chem.
275, 34521 – 34527
53 Jiang, Y. et al. (1999) Prevention of constitutive TNF receptor 1
signaling by silencer of death domains. Science 283, 543 – 546
54 Doong, H. et al. (2000) CAIR-1/BAG-3 forms an EGF-regulated ternary
complex with phospholipase C-g and Hsp70/Hsc70. Oncogene 19,
4385– 4395
55 Sondermann, H. et al. (2002) Prediction of novel Bag-1 homologs based
on structure/function analysis identifies Snl1p as an Hsp70 cochaperone in S. cerevisiae. J. Biol. Chem. 277, 33220 – 33227
TiBS COMPETITION
Are there any talented poets or songwriters in your lab? Did you see ’Coomassie Blues’ in the June issue and feel inspired?
TiBS is running a competition to find the best poem or song relating to TiBS, biochemical sciences or life in the lab. The prize will be
a free subscription to TiBS for one year. The winning and runner-up poems/songs will feature in forthcoming issues of TiBS.
So get writing!
Send your entry to:
Fiona G. Hutton, Editor
Trends in Biochemical Sciences
Elsevier
84 Theobald’s Road
London, UK WC1X 8RR
http://tibs.trends.com